Simpler Navigation for Servers and Operating Systems
Completed: a much simpler Servers and Operating Systems section of the Community. We combined many of the older boards, so you won't have to click through so many levels to get at the information you need. Check the consolidated boards here as many sub-forums are now single boards.
Showing results for 
Search instead for 
Did you mean: 

RAC problem if one node down

Occasional Advisor

RAC problem if one node down


I've 2 nodes HP-UX11i with MC/ServiceGuard & 10g RAC.

When nodeA fails (power or network), listenerA
package failovers to nodeB : ipA is then activated on node B and the command "lsnrctl reload name_listener" is executed on nodeB according to HP-UX doc and metalink notes (247963.1 & 249213.1)

In fact what bothers me is that if I've got more than 2 nodes this can't work : I can't reload the listener on all surviving nodes.

My question is : how can I trigger with MCSG an action on all surviving nodes when 1 node fails (1 package failover) ?

I hope the question is not too vague
Honored Contributor

Re: RAC problem if one node down

I am trying to understand your question.
How many nodes are there in the cluster?? How package failover is configured?? If RAC on nodeA is configured to failover to nodeB, the listener will start on nodeB only. (virtual ip now will be owned by nodeA)

There is no need to start listener on all hosts that are under cluster.

There is no substitute to HARDWORK
Occasional Advisor

Re: RAC problem if one node down

We only have 2 nodes in the cluster but I'm trying to figure out how to make it work if some day we add another node : we should then reload the listener on all surviving nodes (according to metalink note 247963.1).

Right now, if nodeA fails, RAC_instance_pkgA fails, listener_pkgA failover on nodeB. When listener_pkgA starts on nodeB :
- vipA is activated on nodeB
- the command "lsnrctl reload listener_name" is executed on nodeB
Julio Yamawaki
Esteemed Contributor

Re: RAC problem if one node down


As you configured RAC to work with listener failover, if you have more than two nodes, you MUST specify in what order you want listener to be reloaded in case of failover.
Just put in package.conf what nodes and what order you want:
NODE_NAME primary_node
NODE_NAME first_failover_node
NODE_NAME second_failover_node

Esteemed Contributor

Re: RAC problem if one node down

One option is

add apropriate remsh commands to the customer defined run commands section in your package control files so that ,
it does an remsh to other nodes and executes the listener reload commands .

Is this what you are looking for ?

Thanks BL.
Good things Just Got better (Plz,not stolen from advertisement -:) )